Originally Posted by TLa
I came from a reef hobby and in this hobby people tend to use aquarium chiller to keep tank water cool around 78F. These chillers come with automatic temperature control which will turn on itself to specific setting. You might also need a bigger reservoir.
Would this be more efficient that conventional radiator with fan type setting in terms of performance? Has anyone tried it before? I know ppl talk about mini fridge and all but in the aquarium trade, fridge doesn't work either to keep tank water cool.
Very nice discussions on this possibility. I'd like to bring in some numbers and physics... Someone said there was no technical information available by those companies... but I found something on the biggest one that has been submitted in the topic :
The 1/3 HP one
OK so my goal here is to answer to the question : ok it's good to cool an aquarium without that much heat source but will it resist to constant heat from a PC. And a good way to solve these problems is to transfer everything in Energy.
If you look at Pulldown Specs :
The capacity specifications are intended as a guide at a room ambient temp of 86Â°F & heat load of 1.25 W/gallon
* 130 Gallons (27 F Pull down )
* 260 Gallons (18 F Pull down )
* 525 gallons (9 F Pull down )
And the good old equation Q = m*c*deltaT
If you look at the test they made at the bottom of the page, you will see that
in the initial power on of the device, the system pulls down temperature from
78Faranheit to 67Faranheit which correspond to a DeltaT, in celcius, of 6.11 over the course of 105 minutes.
Let's assume there is no heat transfer from the aquarium to the environnement, which is false but in the worst case, this would lead us to underestimate the power of the device. With this assumption, let's evaluate the capacity of the device. How much energy Q did we need to transfer out of the aquarium water to reach a 6.11 deltaT over 105 minutes considering a mass m of 170 kg and a heat capacity of water, which is generally a constant of 4186 J*Kg-1*C-1
The answer is Q = 4347998 Joules
but we didn't transfer this energy in a single shot, we did it over 105 minutes
So if we divide by hours, we will have the power of the device to transfer energy in an hour which gives 2484570 J/Hour, which as I discussed before is an underestimation of the device's power because we assumed that there is no heat transfer from the environnement (lightings, etc) to the aquarium. So the device should be a little bit more powerful than 2484570 J/Hour.
Now at this point of the discussion I'd like to tell you that I'm not a physicist, nor a mathematician, and I have no idea what I'm doing so if you guys know anything about electricity and power consumption and if you want to tell me I'm completely wrong, I won't be offended.
So now the question is does an equipment which pumps 2484570 J/Hour of heat is enough to cool a CPU. I'll investigate that.
Is this calculus realistic ? yes. The device spec says 2620 Btu/H and 2484570 J/Hour = 2356 Btu/H (which shows that our underestimation was of approx 300 Btu/H).
now 2620 Btu/H = 767.3 Watts (or our underestimation = 690 Watts).
As you know, computers use power supplies from 400 to 1000 watts. But these power supplies heat components that you won't include in your circuit : hard disks, dvd readers, etc...
So 767.3 Watts of heat pumpin seems more than enough to cool your CPU and maybe your GPU... Consider that the maximum wattage of your power supply (say 600 watts) is actually never reached...
I think this is really promising. Actually I've come across this thread because I plan to do exactly the same thing in 1 month, and i'll come back with a report... I have a Tai-Chi Watercooling system from Thermataltake with the whole TaiChi case and I'd like to modify it a little bit to include this device, the 1/3 HP. The numbers seem to tell us that's possible.
The only problem thats left is : are these device as good at low temperature. Because for now we investigated the behaviour of the devices at 78-67 Fareinheit, but do they ever offer the option to go lower? Also another thing to consider : 767.3 watts (or 2620 Btu/H) is the heat pumping when the device is always on. So now the % of time the device will be ON depends on how much watts a computer really pumps. I guess it can be from 400 to 600, big maximum, but how much of these watts goes into the CPU and the GPU that we want to cool ? maybe someone can bring us the answer
I just read something that says the whole computer generates 682.80 BTU's / H so the 1/3 HP device is like 3 times what you would need to pump the whole heat generated by the computer, including hard disks and everything
That's good news
Of course this is considering you want to MAINTAIN a temperature. Now what we want is more than maintain, we want to maintain it below ambiant room temperature... So I think the 1/3 HP is a good option.. costs a lot however.